When the efflux of lava from the vent finally ceased, this latter mass will have cooled down with extreme slowness, and been consolidated in a regular and tranquil manner, such as would facilitate the establishment of straight and vertical axes of contraction, and the production of very regular hexahedral columnar concretions, perpendicular, as usual, to the cooling surfaces.* A view of part of this range near the Pont de Baume engraved in the work of Faujas, after due deductions for exaggeration in size and its faulty execution, gives not an inaccurate idea of the disposition of the columns.

It is here that the river, descending both from Burzet and Montpézat, joins the Ardèche; so that, if we suppose their eruptions to have been contemporaneous, or nearly so, it is not at all improbable that the five lava-currents we have described, flowing from as many different points, were once united on this spot, exactly as the rivers were, and still are, whose channels they usurped. In the bed of the river Ardèche, at and for some distance below this point, may be seen in summer, when the stream is inconsiderable, a number of articulations of basaltic columns, in which a nice observer may recognise the mineral characters of the different lava-currents of the tributary valleys. As you follow the course of the river, these columns show themselves less frequently and are more water-worn, till at the distance of a mile or two they are reduced to little more than rounded blocks, and nearly assimilated to the other boulders which cover the dry channel of the river. These basaltic boulders continue to diminish in size as you descend, and few are to be met with as far down the stream as Aubenas so large as a man's head; further on they are reduced to mere pebbles, and are no doubt still more comminuted before the Ardèche

* See Considerations on Volcanos, p. 136.

carries them with it into the Rhône. This observation illustrates the process by which both the basalt and granite that once filled these valleys have disappeared. A wintry flood undermines and detaches a prism of basalt from one of the columnar ranges. The next flood drives it on a few inches; or, if by its form and position it is enabled to roll without much difficulty onwards, a few feet. This operation is repeated year after year, and in the mean time, even when remaining stationary, it is exposed to the immense friction of all the smaller boulders and pebbles which are drifted over it by the ordinary as well as the extraordinary force of the current. By the continuance of this process it is at the same time carried forwards, reduced in size, and brought to approach to a globular form, the most favourable to its transport, in consequence of which the rapidity of its progress along the channel of the river is progressively accelerated, till, diminished to the size of gravel or silt, it is taken into complete suspension, and carried sooner or later in this state into the ocean.

At the foot of the cone of Souillols, near the hamlet of Neyrac, rises a spring strongly impregnated, like that of Jaujac, with carbonic acid gas; and from the bottom of a small hollow close by this gas emanates in such abundance that the exact phenomena of the Grotta del Cane are reproduced, and dogs or fowls may be stifled and revived ad libitum.

6. The last of the cones we have mentioned, called La Coupe d'Ayzac, rises on the ridge of one of the granitic abutments that project from the steep escarpment of the Haut Vivarais. It has a beautiful crater slightly broken down towards the north-west; and from the breach a stream of basalt may be seen to descend the flank of the hill, and turning to the north-east enter the valley of the river Volant, which has subsequently cut it entirely across, and disclosed three distinct storied ranges; the

lowermost very regularly columnar, that in the middle less so, and the upper nearly amorphous, cellular, and with a ragged scoriform surface.* This current, which appears originally to have occupied the whole bottom of the gorge in an extent of about four miles, from the village of Antraigues nearly to that of Vals, where the Volant flows into the Ardèche, has been worn away and carried off on many points by the violence of the torrent. Its relics adhere in vast masses to the granite rocks on both sides, particularly in the receding angles of the valley, sometimes bordering the river for a considerable distance, and reaching the height of 160 feet above it. The lower portion of this bed is very beautifully columnar, the upper obscurely so; this latter has been in parts destroyed, and a pavement or causeway left, formed by an assemblage of upright and almost geometrically regular columns fitted together with the utmost symmetry. The columns, as usual, are always at right angles to the containing surfaces, through which the heat escaped.

Such are the most interesting volcanic remains of the Bas Vivarais; they evidently form the continuation of the line of recent eruptions which we have noticed on the lofty table-land above.

The exceeding freshness of their scoria and the general regularity of their craters would seem to indicate their being of a later date than the others; but this state of superior preservation may to some extent be attributable to their minor elevation and more sheltered position; and the vast amount of excavation already accomplished by the erosive force of the mountain streams along the valleys which feed the Ardèche, since their invasion by currents of basalt, proves the origin of these, however recent in comparison with the beds of the same nature around the Mezen, to belong to an æra incalculably remote

* See the annexed drawing, Plate XV.